Video Annotation in Nurse Home Visits: Methodology Development and Pilot Evaluation

Area of Science:

• Health Informatics
• Clinical Documentation
• Digital Health

Background:

• Home health care (HHC) often misses crucial environmental risk factors in electronic health records (EHRs).
• Video recordings offer a potential solution to capture these overlooked details.
• Standardizing clinical terminologies is essential for integrating diverse data sources.

Purpose of the Study:

• To develop and validate a systematic video annotation methodology for home health care settings.
• To map video-derived annotation codes to standardized clinical terminologies like SNOMED CT.
• To assess the feasibility and practicality of integrating video data with EHRs.

Main Methods:

• Four simulated nurse home-visit videos were systematically annotated.
• Annotation codes were mapped to SNOMED CT and other Unified Medical Language System (UMLS) Metathesaurus terminologies.
• The mapping accuracy and alignment with standardized codes were analyzed.

Main Results:

• A total of 97 annotation codes were identified, with 68% at the data level, primarily environmental observations.
• Over 88% of the identified codes mapped exactly to SNOMED CT, demonstrating high compatibility.
• The methodology proved feasible, with most codes aligning well with standardized clinical terminologies.

Conclusions:

• The proposed video annotation methodology is feasible and practical for home health care.
• This approach can enrich EHR data with valuable environmental observations.
• Future research should explore integration with EHRs for improved patient outcome prediction.